Amazon Tipping Point
The Amazon makes most of its own rain. Roughly half the rainfall over the basin is moisture that the forest itself evaporated days earlier, recycled inland by atmospheric rivers that hop tree to tree. Cut enough trees and the recycling fails. The forest does not gradually thin — it flips, on the timescale of decades, into something closer to a degraded savanna it cannot climb back from.
The case
In 2018, Thomas Lovejoy and Carlos Nobre published the estimate that has anchored the debate since: the Amazon's tipping point sits at 20-25% cumulative deforestation, possibly lower under accelerating climate warming. As of 2024, cumulative forest loss across the basin sits near 17-20%, depending on whose accounting and which boundary. The Brazilian Amazon alone has lost about 20% of its original cover. The window between "current state" and "estimated flip" is narrow enough that it fits inside the noise of measurement.
The mechanism is hydrological, not just botanical. A mature Amazon tree transpires hundreds of liters of water a day. Multiplied across roughly 390 billion trees, this creates the largest evapotranspiration flux on the planet's land surface. Trade winds carry that vapor west and south. Trees downstream catch it, transpire it again, and pass it further. The "flying rivers" feeding São Paulo's rainfall trace back to leaves in Pará.
Remove enough of the upstream forest and the chain breaks. Dry seasons lengthen. Fires ignite in vegetation that evolved without fire. Surviving trees die back. Each dead tree removes a pump from the system. The collapse accelerates the collapse. This is what non-linear means in ecosystem terms: the gradient does not stay gradual.
Where the signal already shows
Southeastern Amazonia, the deforestation frontier, became a net carbon source around 2020 according to Gatti et al. (Nature, 2021). The forest there now emits more carbon than it absorbs. Dry seasons in that region have lengthened by roughly a month over the last 40 years. Tree mortality in long-term monitoring plots has roughly doubled since the 1980s. The southern Amazon is, by several independent measures, already past its local flip.
The wider basin is mixed. The western Amazon still functions as a carbon sink. The eastern Amazon does not. The system is not flipping uniformly — it is unraveling from the edges inward, which is exactly what a hydrologically-coupled tipping system should look like as it approaches its threshold.
What's contested
The 20-25% number is a best-available estimate, not a measured threshold. Lovejoy and Nobre derived it from coupled climate-vegetation models that simplify both the climate and the vegetation. Some modelers argue the true threshold could be 30-40% under cooler climate scenarios, or as low as 15% under high-warming pathways. Others argue the concept of a single basin-wide threshold is wrong — that the Amazon will dieback patchwise, with each sub-basin having its own hydrological budget and its own breaking point.
There is also dispute over reversibility. The standard claim is that the savanna state is a stable attractor — once the forest is gone the climate dries enough that the forest cannot regrow. Paleoclimate evidence from the last glacial maximum, when the Amazon was drier and patchier, complicates this. The forest did come back. The question is whether it can come back when humans are simultaneously warming the planet by 2-3°C.
Why this has to do with other realms
The Amazon tipping point is the cleanest worked example of a concept tipping point in earth systems: a self-reinforcing feedback loop with a threshold whose location is uncertain but whose existence is not. The economics rhyme with bank runs and currency crises — slow, then sudden. The mechanism rhymes with concept phase transition in physics. The political economy rhymes with concept tragedy of the commons: each Brazilian rancher's clearing is rational at the margin, catastrophic in aggregate, and the aggregate is what decides whether the system flips.
It also sits adjacent to concept planetary boundaries — the Stockholm Resilience Centre's framework treats land-system change and biosphere integrity as two of the nine boundaries the Amazon sits at the center of.
An open question
If the southern Amazon has already crossed its local threshold and is now a carbon source, has the basin-wide tipping point already begun, with the rest of the system following on a 20-40 year lag? Or is the eastern dieback a regional phenomenon that the western Amazon's intact moisture pump can compensate for? The honest answer in 2026 is that the satellite record is not yet long enough to tell.
Key sources
- Lovejoy & Nobre (2018), Science Advances, "Amazon Tipping Point" — the paper that named the 20-25% figure.
- Gatti et al. (2021), Nature, "Amazonia as a carbon source linked to deforestation and climate change" — the southeastern carbon-source finding.
- Boulton, Lenton & Boers (2022), Nature Climate Change, on resilience loss in the Amazon since the early 2000s.
- to verify: MapBiomas annual deforestation accounting for cumulative-loss figures by sub-basin.
- to verify: Nobre's 2023 follow-up work updating the threshold estimate under current warming.
Further reading
- The Future of the Amazon (Nobre, in interviews 2022-2024) — the clearest articulation of why moisture recycling is the load-bearing mechanism.
- The INPE PRODES dataset — Brazil's official deforestation monitor, public, annual, going back to 1988.
- Carlos Nobre on the Mongabay Newscast — accessible long-form on the tipping point debate.
- Earth Transformed by Peter Frankopan (2023) — situates the Amazon in the longer history of human-forced ecological flips.
See Also
- concept tipping point — the general theory the Amazon is the worked example of.
- concept planetary boundaries — the framework Amazon sits inside.
- concept phase transition (cross-realm: physics) — same mathematical shape, different substrate.
- concept tragedy of the commons (cross-realm: economics) — why rational local actors produce irrational aggregate outcomes.
- concept carbon cycle — what flips when the Amazon flips.
- person carlos nobre — the climatologist whose career is this question.